By: Alain Geenrits, Blue Medora
Customers are always looking to reduce admin effort and TCO of their virtual environments. In a typical setup, VMware admins have to look after the X86-servers their vSphere hosts are running on, some form of storage area network – SAN or NAS – and the network in between all those elements. A data center might have generations of products from diverse vendors, acquired over the years. This can make provisioning or repurposing slow.
A few years ago vendors started introducing ‘converged’ infrastructure in the market. In this Lego-style approach, all the building blocks for your infrastructure are standardized, validated and tested for your VMware environment. By using these standard building blocks it becomes easier to do administration, to install software updates and repurpose or expand parts of your environment.
The two prevalent families are:
- NetApp FlexPod: Cisco UCS servers, Cisco Nexus switches and almost any NetApp storage, building on Data Ontap (Clustered Ontap).
- EMC VCE: Cisco UCS servers, Cisco Nexus switches and EMC storage arrays.
Naturally this standards-based approach benefits from an equally centralized management environment. Full stack visibility takes on a new meaning here. For FlexPod we had individual solutions at Blue Medora: the management packs for NetApp, Cisco UCS and Cisco Nexus enable you to manage all building blocks of the converged infrastructure in your VMware vROps management platform. Our FlexPod bundle gives you dashboards and reports for a complete solution.
Figure 1: FlexPod Heatmap Overview
Last month, Blue Medora released a Management Pack for VCE Vblock. This plugin began its life as a solution from VMware. Unlike Flexpod, this solution is not a bundle. VCE provides a central management interface called Vision Intelligent Operations software. We interface with that API to manage vBlock models 100 to 740 with vRealize Operations. For more information see the user guide.
Figure 2: VCE Vision Inventory
We bring you a number of dashboards out of the box highlighting key components and metrics and providing smart alerts. Of course we also use the unique ability in vROps to detect relationships between objects, critical in a converged infrastructure, where we build precisely on these interconnected blocks.
Figure 3: VCE Object Relationship Example
The converged model is now evolving into what we call a hyper converged architecture. This is based on what VMware calls the Software-defined datacenter: all functions run in software on “vanilla” hardware, the network routers as well as the storage arrays. VMware started building this with NSX network virtualization. Instead of an expensive and complex network in your datacenter, all networking is done in a sort of hypervisor for the network, hence the brilliant name NSX (building on ESX for servers). All you need is standard X86 servers and network switches.
VMware also introduced vSAN, software-defined storage or a hypervisor for storage. It lets a standard server with hard disks and flash drives act as a very fast storage appliance. You can see the writing on the wall: the complete converged stack we talked about before can now be implemented in software on standard servers. It brings closer the vision of running data centers like Facebook and Google: rows and rows of standard – and redundant – servers with all functions implemented in software. Lightning fast (re)deployments become the norm, easy admin through automation and a very low TCO. VMware launched this bundle as HCS (Hyper-Converged Software). VCE launched an appliance called VxRail based on this technology, and also launched VxRack based on EMC’s own ScaleIO software-defined storage.
There are of course other players on this hyper converged market and Nutanix seems to be the most popular one. They just acquired PernixData, also a player in storage virtualization. Nutanix approaches the convergence question with their own technology based on the Acropolis hypervisor. They can run VMware workloads. The hosts and storage again run virtualized on standard hardware. A lot of customers are adopting this platform and asked Blue Medora to create a solution to manage from vROps, and the week before VMworld the Management Pack for Nutanix was released!
Figure 4: Nutanix Topology Widget Showing Relations
The solution works on Nutanix hardware versions NX1000+ and connects to the Nutanix REST API v1+ through a Prism Element or Prism Central IP address. The difference being the first one returns vDisk resource configuration details, but only in the connected cluster, while the later does not return this resource, but can see multiple clusters.
Figure 5: Topology Focus on Nutanix Host with KPI Data
We bring smart alerts and capacity planning to the table, allowing you full control of your Nutanix environment through vRealize Operations. For more info contact us on our website or Twitter @BlueMedora or email as at firstname.lastname@example.org.
Figure 6: Nutanix Overview Dashboard
In this blog post we have talked about converged and hyperconverged architectures in the market. Their aim is to reduce admin time and complexity by using interchangeable building blocks. Blue Medora delivers management solutions in vRealize Operations for the most important players in this field. Whatever solution you implement, you will be able to manage it with vROps! If you have suggestions for other solutions we should work on, please contact us.